detection for still running qemu instance in worker only works with "--no-cleanup"
See #55505#note-7 : openQA has a line "[error] [pid:…] A QEMU instance using the current pool directory is still running (PID: …)" which should be printed e.g. in cases like #55505 but this does not happen in production workers running with pool dir cleanup.
Apparently the worker cleans up the pool directory so that the qemu pid can not be read out from local files. Might be a regression introduced with recent worker rework.